Misconceptions about the UK system of corporate taxation
-
Posted by David Massey
- Posted on 3/28/2023 9:03:35 AM
I believe the results reported by this paper may be unreliable. The main problem is that it is based on fundamental misunderstandings of the system of the corporation tax system in the UK. In particular, it fails to appreciate the effect of group structures on the tax returns of individual companies. Membership of a group has implications for the “size” of a company, its rate of tax and the due dates for payment.
The paper attempts to identify its treatment group by looking at a company’s turnover (p. 366) taken from Box 1 of the tax return form CT600 (Table B2, page 386), or its assets (p. 376) as reported in the statutory accounts. However the size of a company is not determined at a company level rather it depends on the consolidated results of the whole undertaking to which the company belongs. [SS 47-49 CAA 2001]. We are told that 83% of treated companies are part of a larger undertaking (p. 384). I would not be surprised if many of these did not actually qualify for First Year Allowances (FYAs) as SMEs (Small or Medium Enterprises).
Despite not being SMEs, these larger companies may still have correctly claimed FYAs. The paper mis-states the availability of FYAs. It asserts that they were available only to SMEs (p. 364). However, FYAs on some assets (such as low CO2 emission cars and “Green” technologies) could be claimed by all companies. The CT600 does not distinguish between the two types of FYAs. I am not a STATA user and am happy to be corrected here, but it appears that companies may have been incorrectly excluded from the analysis where they may have (correctly) claimed FYAs which were not expected on the assumptions made by the paper. And that investments qualifying for FYA has not been included for all companies or for all years. I think the effect of these may have been to overstate the increase in investment in the period examined.
The paper determines the marginal rate of tax by comparing the “taxable income” for a period to the rates of tax reproduced on p.367 (Table 3). But the marginal rate of tax can NOT be calculated this way.
The rate of tax depends not simply on the taxable profits reported at Box 37 of the CT600 (p. 386) but rather on what we currently call “augmented profits” – Box 37 plus Box 38 (franked investment income) – which includes any dividends received from outside the group.
The rate of tax also depends on the number of “associated” companies. The rate band thresholds are reduced proportionately for each associate a company has. So, for example, if a company has one associated company, its thresholds will be reduced to ½ of the amounts given in Table 3. It will have become liable to the full (30%) rate of corporation tax if its profits exceeded £750k, rather than the £1.5m given by the table. The thresholds for the other rates would be similarly reduced (to £150k, £25k, and £5k). If a company had 4 associates then the thresholds would be reduced to 1/5 of the values given in the table (£300k, £60k, £10k, and £2k). There are a number of ways in which companies can be associated outside of a group structure, but companies in the same group will always be associated. It is reported on p.384 that almost 90% of companies were determined to be members of a group. There is a fair chance that many companies will have been placed in the wrong tax band.
The number of associated companies are included in the CT600 at Box 39, or Boxes 40/41 where the return covers a period which straddles 31 March. Box 39 is included within the HMRC Dataset but boxes 40 and 41 are not. (See the attached link). The paper reports that about 2/3 of their sample had December year ends (page 379 fn.18). This would imply that the associated company information is likely to be missing from the dataset for a substantial number of companies.
In Section V, the paper reports that it uses differences in tax payment dates to test for “cash flow effects” (p. 381ff). It asserts that companies with taxable profits below £1.5m will pay tax 9 months after their accounting year, while those with profits above that amount will pay by quarterly instalments, all of which are due well in advance of the usual date. However, that £1.5m threshold applies in the same way as it does for determining the rate of tax in that the profits to be used are the ”augmented profits” and the threshold is apportioned based on the number of associated companies. There are some complications for companies going over the threshold for the first time and in the date at which the associated companies are counted, and for companies with very small profits. This needn’t have caused a problem as companies paying by instalments identify themselves by checking Box 95, but again this appears to have been an item that was not included within the dataset.
I have doubts that the HMRC Datalab dataset can be used to answer the questions the paper has set itself.
The AEJ:EP has published other papers which share some of the weaknesses of this one:
Devereux, Michael P., Li Liu, and Simon Loretz. 2014. "The Elasticity of Corporate Taxable Income: New Evidence from UK Tax Records." American Economic Journal: Economic Policy, 6 (2): 19-53. DOI: 10.1257/pol.6.2.19
Guceri, Irem, and Li Liu. 2019. "Effectiveness of Fiscal Incentives for R&D: Quasi-experimental Evidence." American Economic Journal: Economic Policy, 11 (1): 266-91. DOI: 10.1257/pol.20170403
I will try to comment on those papers as soon as I can, where I will also flag up some other challenges that may also impact on this paper.
David Massey CTA (Fellow) ATT 675818@soas.ac.uk
Misconceptions about the UK system of corporate taxation
The paper attempts to identify its treatment group by looking at a company’s turnover (p. 366) taken from Box 1 of the tax return form CT600 (Table B2, page 386), or its assets (p. 376) as reported in the statutory accounts. However the size of a company is not determined at a company level rather it depends on the consolidated results of the whole undertaking to which the company belongs. [SS 47-49 CAA 2001]. We are told that 83% of treated companies are part of a larger undertaking (p. 384). I would not be surprised if many of these did not actually qualify for First Year Allowances (FYAs) as SMEs (Small or Medium Enterprises).
Despite not being SMEs, these larger companies may still have correctly claimed FYAs. The paper mis-states the availability of FYAs. It asserts that they were available only to SMEs (p. 364). However, FYAs on some assets (such as low CO2 emission cars and “Green” technologies) could be claimed by all companies. The CT600 does not distinguish between the two types of FYAs. I am not a STATA user and am happy to be corrected here, but it appears that companies may have been incorrectly excluded from the analysis where they may have (correctly) claimed FYAs which were not expected on the assumptions made by the paper. And that investments qualifying for FYA has not been included for all companies or for all years. I think the effect of these may have been to overstate the increase in investment in the period examined.
The paper determines the marginal rate of tax by comparing the “taxable income” for a period to the rates of tax reproduced on p.367 (Table 3). But the marginal rate of tax can NOT be calculated this way.
The rate of tax depends not simply on the taxable profits reported at Box 37 of the CT600 (p. 386) but rather on what we currently call “augmented profits” – Box 37 plus Box 38 (franked investment income) – which includes any dividends received from outside the group.
The rate of tax also depends on the number of “associated” companies. The rate band thresholds are reduced proportionately for each associate a company has. So, for example, if a company has one associated company, its thresholds will be reduced to ½ of the amounts given in Table 3. It will have become liable to the full (30%) rate of corporation tax if its profits exceeded £750k, rather than the £1.5m given by the table. The thresholds for the other rates would be similarly reduced (to £150k, £25k, and £5k). If a company had 4 associates then the thresholds would be reduced to 1/5 of the values given in the table (£300k, £60k, £10k, and £2k). There are a number of ways in which companies can be associated outside of a group structure, but companies in the same group will always be associated. It is reported on p.384 that almost 90% of companies were determined to be members of a group. There is a fair chance that many companies will have been placed in the wrong tax band.
The number of associated companies are included in the CT600 at Box 39, or Boxes 40/41 where the return covers a period which straddles 31 March. Box 39 is included within the HMRC Dataset but boxes 40 and 41 are not. (See the attached link). The paper reports that about 2/3 of their sample had December year ends (page 379 fn.18). This would imply that the associated company information is likely to be missing from the dataset for a substantial number of companies.
In Section V, the paper reports that it uses differences in tax payment dates to test for “cash flow effects” (p. 381ff). It asserts that companies with taxable profits below £1.5m will pay tax 9 months after their accounting year, while those with profits above that amount will pay by quarterly instalments, all of which are due well in advance of the usual date. However, that £1.5m threshold applies in the same way as it does for determining the rate of tax in that the profits to be used are the ”augmented profits” and the threshold is apportioned based on the number of associated companies. There are some complications for companies going over the threshold for the first time and in the date at which the associated companies are counted, and for companies with very small profits. This needn’t have caused a problem as companies paying by instalments identify themselves by checking Box 95, but again this appears to have been an item that was not included within the dataset.
I have doubts that the HMRC Datalab dataset can be used to answer the questions the paper has set itself.
The AEJ:EP has published other papers which share some of the weaknesses of this one:
Devereux, Michael P., Li Liu, and Simon Loretz. 2014. "The Elasticity of Corporate Taxable Income: New Evidence from UK Tax Records." American Economic Journal: Economic Policy, 6 (2): 19-53. DOI: 10.1257/pol.6.2.19
Guceri, Irem, and Li Liu. 2019. "Effectiveness of Fiscal Incentives for R&D: Quasi-experimental Evidence." American Economic Journal: Economic Policy, 11 (1): 266-91. DOI: 10.1257/pol.20170403
I will try to comment on those papers as soon as I can, where I will also flag up some other challenges that may also impact on this paper.
David Massey CTA (Fellow) ATT 675818@soas.ac.uk